ChatGPT, Claude and Grok go down together in an unusual simultaneous outage, with Gemini also affected

On Thursday morning three of the leading artificial intelligence chatbots went down at once: OpenAI's ChatGPT, Anthropic's Claude and xAI's Grok, according to reports collected by the monitoring platform Downdetector. Problems were also logged with Google's Gemini, although the company has not issued an official notice confirming an outage.

What stands out about the episode is not the outage itself —service disruptions on AI platforms are relatively frequent— but the fact that all three services went down at the same time, just as companies and users are relying more and more on these tools for their daily work. Each company gave slightly different details about the scope of the problem: Anthropic's status page flagged specific disruptions in Claude's Opus 4.8 and Opus 5 models, clarifying that its other models had already "recovered to baseline error rates." OpenAI, for its part, reported "elevated errors" in both ChatGPT and Codex. xAI confirmed that Grok "is experiencing issues" and that its team was working on a fix.

The most significant detail in the article is the possible common thread: Microsoft Azure, which serves as the cloud infrastructure provider for all three services, was also logging outages of its own at that very moment. Some specialist outlets suggested that the Azure failure could be behind, at least in part, the chain of AI model outages. If confirmed, that hypothesis would highlight the extent to which the generative AI industry —despite the apparent competition among OpenAI, Anthropic, xAI and Google— in practice shares a cloud infrastructure heavily concentrated in a handful of providers (Azure, AWS, Google Cloud), turning a one-off failure at a single provider into a systemic risk capable of taking down direct competitors simultaneously.

The original article is a breaking news piece subject to updates, so it does not yet offer a confirmed cause, the duration of the incident or figures on the number of users affected. The information should therefore be treated as preliminary until the companies involved —OpenAI, Anthropic, xAI, Google and Microsoft— publish postmortems or more detailed confirmations about the technical origin of the blackout.